Scale-invariant object identification for jointly addressing issues in moving oject and background
In the process of moving target detection, moving objects and their shadows are often extracted both. It leads to the failure of moving target positioning, segmentation, tracking and classification in the moving-object detection. Thus the shadow removal is an important task in the intelligent monitoring system. In the process of shadow removal, using color features to detect shadow is easy to cause that some parts of the prospect with the same color as the shadow may be mistakenly removed, so that have an impact on the connection and integrity of the prospect. In this paper, in order to obtain a complete prospect as well as remove shadow, we propose a moving target detection algorithm based on shadow removal and prospect reconstruction. Our algorithm can effectively restore the prospect which is mistakenly removed as shadow, and the experimental results confirm our algorithm.